The Data You Don't Have Will Cost You More Than You Think 💰
I had one of those conversations recently that stays with you. The kind where someone says something that reframes how you see an entire industry.
One of our PIM customers — a UK manufacturer with hundreds of products across multiple categories — had just been through a supplier onboarding exercise with a major online retailer. We're talking about a household name, the kind of company you'd assume has every process nailed down, every system humming, every piece of product data flowing seamlessly from supplier to shelf.
They don't.
The Assumption That Fell Apart 🤔
Our customer had spent about two weeks setting up their product data properly in the PIM. Structured families, configurable parent-child relationships, automated descriptions, dynamic PDF datasheets — the full picture. When the retailer sent over their supplier data template, our customer expected something sophisticated. What arrived was… basic. Missing categorisation levels. No configurable product structure. Fields that didn't correspond to how products were actually displayed on the retailer's own website.
When our customer pointed this out — gently explaining that kitchen products are essentially the same item in different widths and colours, and that you'd normally model this with configurable parent-child relationships — it took quite a while to get the concept across.
The word "manual" came up more than once. A lot of their merchandising, it turned out, was being done by hand. People editing website listings directly. No centralised product data system feeding the site. Just teams of people, presumably divided by product category, each manually maintaining their corner of what must be a catalogue running into the millions of SKUs.
As our customer put it: "You just assume these massive companies are well-oiled machines. I don't think they are."
Two Weeks vs. Two Departments ⚡
Here's where the contrast gets sharp. Our customer had invested two weeks getting their data into the PIM properly. Not because the retailer demanded it — in fact, the retailer's template was so basic that the business was going to just fill it in manually and send it back. Our customer stopped them.
"Hang on," they said. "This doesn't even match how products are set up on their site. Let's put it in the PIM properly, because when they change their mind — and they will — we can react in hours instead of weeks."
They were right. The retailer did come back with revised requirements. And because the data was structured, the revised output took hours. Not days, not a new round of meetings and manual re-entry. Hours.
Then our customer showed up to a face-to-face meeting with PDF datasheets generated dynamically from the PIM. Product summaries that would have taken a design team days to lay out manually, created in minutes from live data. The retailer's reaction? "Oh yeah, that's good."
No questions about how it was done so fast. No curiosity about the tooling. Just acceptance — because they didn't have the frame of reference to realise how impressive it was.
That's the thing about operating manually at scale. You don't know what you don't know. You can't see the gap because you've never stood on the other side of it.
The Compounding Problem 📈
This story isn't really about one retailer's manual processes. It's about what happens next.
We're entering a period where AI is going to transform how businesses use product data. Not in some theoretical, five-years-from-now way — right now, today. Language models can already generate product descriptions, translate content across languages, create presentations, write export scripts, and answer natural language questions about product catalogues. And they're getting better every quarter.
But here's the catch: AI is only as good as the data behind it. Garbage in, garbage out — or as our customer more colourfully put it, the industrial-strength version of that phrase.
If you've got a well-structured PIM with clean, validated, centralised product data, then AI becomes a multiplier. You can have a conversation with it: "Create a presentation with these four products, our logo, and the key specifications." Done. "Translate these 500 descriptions into German and French." Done. "Write me an export script that reformats our data for this retailer's new template." Done.
If you haven't got that foundation? The AI has nothing to work with. You're asking it to build a house without bricks.
And this is where the compounding effect kicks in. The companies that have their data organised now won't just be more efficient today — they'll be the ones who can actually leverage AI when it matures. The ones still doing things manually won't just be behind; they'll be behind and unable to catch up quickly, because getting product data organised across an entire business is not a trivial exercise. It takes weeks of decisions about taxonomy, attribute structures, approval workflows. It requires input from across the company. It's fundamentally a human coordination problem that no AI can shortcut.
As our customer noted: "When they do realise they've fallen behind, you're already ahead. And it takes them too long to catch up." ⏳
The Real Value of a PIM in 2026 🚀
There's been a quiet shift in what a PIM system actually is. It used to be about organising data and publishing it to channels — e-commerce platforms, marketplaces, PDF catalogues. That's still true, but it's become the minimum.
The real value now is that a PIM is the foundation layer for everything that comes next. It's the single source of truth that makes AI integration possible. It's the structured data that allows you to react in hours when a major retailer changes their requirements. It's the knowledge base that lets you generate dynamic content — datasheets, presentations, translations — without throwing people at the problem.
Think of it this way: every business has product data. The question is whether that data is organised, validated, and accessible through APIs, or whether it's scattered across spreadsheets, email threads, and the heads of people who might leave. The first scenario is ready for the AI age. The second is a ticking clock.
Our customer gets this. They told me the PIM is becoming "a real pivotal part of how we do things" and that they see it as central to their commercial strategy going forward. Not because of what it does today — although the speed advantages are already paying for themselves — but because of what it enables tomorrow.
The retailer, meanwhile, has all the resources in the world. They've got augmented reality on their product pages, for goodness' sake. But underneath that shiny front end, there's a manual process that can't scale, can't adapt quickly, and certainly can't plug into the AI capabilities that are about to reshape e-commerce.
You can have the best AI in the world. But if you haven't got the data, it's got nothing to think about. 🧠
Pat Violaris is the Managing Director of OneTimePIM, a Product Information Management system built by people who've been wrestling with product data since 1992. He has a PhD in Expert Systems, which is what they called AI before it was cool.